
Meta-Apply Patch Lander
Lands staged self-optimization patches into the corpus after a fresh review and your explicit approval.
- Stars
- 18,482
- Licence
- MIT
- Files
- 1 file
What it does
This skill is the only step that turns a proposed change to the ARIS research harness (its own skills and corpus files) into a real edit. Earlier read-only skills stage candidate patches in a pending folder; when you type /meta-apply and name patch numbers, this skill runs a fresh reviewer on each staged diff, refuses any patch the reviewer rejects, backs up the target file, applies the change with Edit or Write, and records a provenance stamp. The developer's explicit invocation is the approval step.
Great for
- Developers maintaining the ARIS auto-research harness who have reviewed a /meta-optimize REPORT and want to accept specific patches
- Maintainers who need an audit trail showing whether a landed change was reviewed by a same-family or a different-family model
Skip it if
- Anyone wanting to apply code changes to an ordinary project: it only handles staged ARIS self-modification patches under .aris/meta/pending/
- Unattended or automated runs: it is designed to be human-invoked only, and a reviewer KILL cannot be overridden, so blindly running 'all' without reading the REPORT risks changing the harness in unwanted ways
Try asking
- “/meta-apply 1,3 after reading the REPORT from /meta-optimize”
- “/meta-apply all to land every staged patch that passes the fresh review”
- “land the staged patches”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the MIT licence and an attribution note.
meta-apply.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip meta-apply.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip meta-apply.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Read the staged manifest and REPORT before naming patch numbers; the producer's pre-screen is advisory only and the fresh review decides.
- Expect same-family reviews to produce only provisional stamps, which do not enable future auto-curation.
- Check the backup folder under .aris/meta/backups/ if you need to roll a change back.
What's inside
- SKILL.md
Redistributed unmodified from wanshuiyin/Auto-claude-code-research-in-sleep under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.